It is known to use a server for downloading data to a plurality of devices. For example, a server may be in charge of downloading data related to Telecom subscriptions in a fleet of Universal Integrated Circuit cards (UICCs). Generally, such a server comprises a main data storage machine which is associated with a backup machine. The backup machine allows to restart the service under acceptable conditions when a serious incident occur at the active server. Preferably, the main data storage machine and the backup machine are arranged so as to provide geographic redundancy.
Since the backup machine is intended to contain at least one copy of all data considered worth saving, the data storage requirements can be significant. Organizing this storage space and managing the backup process can be a complicated undertaking. Since real time synchronization is very costly, many servers are designed with a synchronization latency which can last up to several minutes. Such servers can face synchronization errors when the main data storage machine goes down since actions performed on the main data storage machine side are not yet reflected on the backup machine side.
There is a need for enhancing servers designed with a synchronization latency between the main data storage machine and its associated backup machine.